Lessons from IBM Watson Health with Doug Meil, Software Architect and Author

Summary

---

Lessons from IBM Watson Health with Doug Meil, Software Architect and Author

Becker’s Healthcare Podcast | September 16, 2025
Guest: Doug Meil, software architect & author of “The Rise and Fall of Explorys and IBM Watson Health”
Host: Scott Becker

---

Episode Overview

In this episode, Scott Becker interviews Doug Meil, founding engineer of Explorys and author of "The Rise and Fall of Explorys in IBM Watson Health." Doug gives an insider perspective on the ambitious vision, dramatic failures, and lasting lessons from IBM Watson Health—a much-hyped healthcare and AI venture. They explore why Watson Health failed, what could have been done differently, and the enduring product and business lessons for the healthcare tech sector.

---

Key Discussion Points & Insights

1. The Rise and Acquisition

• Explorys' Backstory: Founded in 2009, Explorys was acquired by IBM in 2015 as part of Watson Health's rapid expansion.
• IBM’s Ambitions: Watson Health was positioned as a "moonshot" (01:29), riding high on IBM’s Jeopardy! win and broader AI hype.

2. The Shadow of Jeopardy!

• Doug describes the overhang of IBM’s 2011 Jeopardy! victory—brilliant in its own context, but ill-prepared for healthcare's complexity.
• Quote:

  “It was a really good Jeopardy-winning machine... But applications to other industries are going to take time. And that’s something, that’s a lesson that I think anybody in product software will understand. Unfortunately, I don’t think that IBM’s leadership fully ingested that message.”
  — Doug Meil, (02:22)

3. Why Did Watson Health Fail?

Pressure and Expectation Management

• The 2010s were a tough decade financially for IBM—there was immense pressure for quick, spectacular wins (03:37).
• IBM acquired multiple companies (Explorys, Phytel, Merge Healthcare, Truven, Cúram) and attempted to blend many disparate business lines and technologies under one umbrella (04:41).

Product Strategy Missteps

• IBM's "offering management" (product management) forced all teams to focus on new products only, rather than enhancing successful legacy ones.

• Quote:

  “Offering management said we will give no priority to existing products. Everything must be a new offering. And that was a very strange message to the customer bases... On a one to two year time frame, it was just suicidal.”
  — Doug Meil, (05:37)

• IBM failed to communicate what “Watson” actually was: “Cognitive Computing” was a vague marketing term.

• Quote:

  “It was hard to get a definition of what Watson exactly was… If you were like a business person… wildly inflated expectations of ‘this thing can do anything’.”
  — Doug Meil, (07:17)

Technology Lock-In and Internal Barriers

• IBM wanted all Watson products to run exclusively on IBM’s hardware and cloud, hindering integration with acquired companies’ products (07:40).
• Years later, products like WatsonX are now deployed on AWS and Azure, proving Doug and his peers’ arguments valid.

4. Industry Context and Market Readiness

• Scott and Doug agree that the market wasn't ready to grasp AI in healthcare at the time; now, with ChatGPT and others, there’s context for what’s possible (09:02).

5. Lessons for Today's Healthcare Tech

• Customer Centricity:

  “For product development, customers count, customers count and keep iterating. Velocity wins.”
  — Doug Meil, (09:54)

• Stagnation due to “endless planning and replanning” led to product failure and customer attrition.

• Host Reflection:

  “You end up not building something your customers really want unless you’re really, really engaged in talking to them… iterate till you get a product fit that people actually want.”
  — Scott Becker, (10:20)

• Adage: It’s easy to say, hard to do, but this advice never goes out of style.

6. Acquisitions, Integration, and Organic Growth

• Integration is often underestimated.

• Organic skills are critical for merging acquisitions—otherwise, “buying growth” backfires.

  “The act of integrating those acquisitions into an integrated portfolio requires organic thinking and organic skills.”
  — Doug Meil, (11:04)

• Both Explorys and Phytel had organic growth pre-acquisition, but a lack of healthcare background and product experience at the parent level undermined their potential (12:21).

---

Notable Quotes & Memorable Moments

• On Jeopardy’s Symbolism:
  "That cast a long shadow, but probably for some of the wrong reasons.”
  — Doug Meil, (01:59)
• On Product Priorities:
  "Everything must be a new offering... On a one to two year time frame, it was just suicidal.”
  — Doug Meil, (05:51)
• On Customer Voice:
  “If you’re not talking with your customers, somebody else will and you will be replaced.”
  — Doug Meil, (10:01)
• On Acquisitions vs. Organic Growth:
  "Acquisitions happen, but the act of integrating those acquisitions... requires organic thinking and organic skills.”
  — Doug Meil, (11:04)
• On Execution:
  “There’s definitely a lot wanting in execution.”
  — Doug Meil, (12:46)

---

Key Timestamps

• 00:38 – Doug Meil introduces himself and Explorys’ origin
• 01:29 – IBM’s “Healthcare moonshot” and Jeopardy! legacy
• 02:22 – Lessons from Jeopardy: domain adaptation doesn’t happen overnight
• 04:41 – The crowded, chaotic Watson Health acquisition spree
• 05:37 – The “everything must be new” product misstep
• 07:17 – Defining (or not) “Watson”: the impact of vague marketing
• 09:54 – Top lessons: customer feedback, velocity, iteration
• 11:04 – Acquisitions, integration, and why organic growth matters

---

Final Takeaways

• IBM Watson Health’s downfall was not for lack of vision, but failed execution: Overambitious timelines, underappreciation of integration complexity, customer neglect, and unclear product strategy all contributed.
• Customer feedback and continuous iteration are essential for product success, especially in healthcare technology.
• Successful M&A in tech requires both organic growth and the skills to genuinely integrate new parts—buying businesses alone is never enough.
• Realistic timelines and clear expectations should govern any attempt to “moonshot” major innovations in complex, regulated spaces like healthcare.

---

Book Mentioned:
The Rise and Fall of Explorys and IBM Watson Health — by Doug Meil (09:41, 12:59)

---

For listeners seeking a candid, practical look at the dynamics of tech, healthcare, and AI innovation, this episode is a must-hear.

Transcript

[A]

This is Scott Becker with a special episode of the Becker Healthcare and Becker Business and Private Equity podcast. We're joined today by a brilliant software engineer, Doug Mel. And Doug has written a book called the Rise and Fall of Explorers in IBM Watson Health. And Doug was right in the middle of this for a decade or more. It's a fascinating story with IBM Watson Health. For those of you that don't remember, it was the hottest thing in healthcare 10, 15 years ago and slowly sort of fell away. Doug's going to tell us a of the story, a little bit about the book they wrote. Doug, can you take a second to introduce yourself and tell us about the book?

[B]

Sure thing. Thank you, Scott. So my name is Doug Neal. I am from Cleveland, Ohio. I've had 30 years in the software and data space and I was the founding engineer of Explorist in 2009, which was acquired by IBM in 2015. And then unfortunately things didn't work out as planned. And so the book represents about a decade of my life. It was, I cared very much about the healthcare mission and we had a very strong engineering team. And yeah, it's just a shame that things didn't work out like everybody had hoped.

[A]

No. And at one point IBM Watson Health was sort of the darling of the tech industry, the software healthcare industry. And so what happened? Why did it fail and what maybe could have been done differently?

[B]

That's a great question. So IBM Watson Health got a lot of coverage both when the division started. It was called a moonshot by IBM itself. This is going to be, you know, IBM's Healthcare Healthcare moonshot. So there was a lot of fanfare on it and we were looking forward to it. We were really excited after we got acquired. You also have to look back to the shadow of jeopardy. So Jeopardy, IBM's, you know, well known in 2011 where they, you know, they, they beat Ken Jennings, they, they won Jeopardy. That cast a long shadow, but probably for some of the wrong reasons. I also ran a, a tech meetup in Cleveland called the Cleveland Big Data Meetup. I had one of the systems engineers from Jeopardy. Talk at my meetup. He did an extra long presentation. It was terrific. One of the, the things he was talking about in 2015 when he did the presentation were, was how long it took to actually win a Jeopardy. They had many, many, many, many iterations of performance tuning and they had to write special software to understand the rules of the betting and how to buzz in quickly. And what you got from, from that presentation was, wow, this was a really good Jeopardy. Winning machine. It was a great jeopardy. Winning software. But applications to other industries are going to take time. And that's something, that's a lesson that I think anybody in product software will understand. Unfortunately, I don't think that IBM's leadership fully ingested that message, so to speak. They saw the marketing potential and the hype, but they forgot that investment in new industries, especially ones as complicated as healthcare will take a lot of time. And they're not easy problems. Problems.

[A]

Thank you. And what could have been done differently to, to make this all work? They were sort of like lost with the jeopardy. Example. What was going on there. Couldn't quite figure out how to translate it all to healthcare. Tell us what they could have done differently.

[B]

That's fair. So also in context, the 2010s were a, were a tough decade for IBM. It's easy to forget that now, but IBM was going through quarter over quarter decline and I think it also made them a bit desperate for wins that were just too quick and they were trying to optimize too many knobs at one time. You know, we need billions and billions of revenue but we also want to be profitable. But we also want to, you know, promote every, every IBM tech available. And those are, those are all tough to optimize all at the same time. Does that make sense?

[A]

100%. It's like trying to do too many things at one time. Total.

[B]

Totally.

[A]

Totally. And IBM people don't remember IBM is rocking and rolling again, but they went through a couple long dry periods post the Lewis Gerstner period, a different period where they were trying to figure out product versus consulting and they've gone through a number of different rough spells and now they're doing great again. But no, your point on that for people that don't know the history is really helpful.

[B]

And so when we, when Explorist was acquired in 2015, it was right in the middle of that decade. So the revenue pressure, the profitability pressure was front and center not just for IBM as a whole, but also for our division. And to remind folks of the context of Watson Health, Watson Health bought Exploris in April of 2015 and also bought Fitel in April of 2015 it bought Merge Healthcare primarily in radiology and diagnostics. In the fall of 2015 it bought Truven in the spring of 2016 and it also merged in a company called Curum which had been acquired a couple of years earlier as well as also that did social services software. Also the Watson for oncology effort which predated Watson Health division was thrown in as well as several of the research efforts. So you had, there wasn't a single Watson health experience, if that makes sense. There were a lot of things happening and at least five different acquisitions, external acquisitions, plus internal projects and trying to rationalize all of that. It's a big effort. And so one of the things that offering management did and offering management is something that is what IBM calls whatever everybody else in the industry calls product management. But IBM's offering management said we will give no new, we will give no priority to existing products. Everything must be a new offering. And that was a very strange message to the customer bases of the acquired companies such as Explorus, Vital and Merge and Kyrum where they said it's all new products. And that is an aggressive message on a five to seven year timeline, on a one to two year time time frame. It's, you know, it was just suicidal. And the market analysis showed that it reflected that of just people were being like, I don't know where you guys are going 100%.

[A]

And again against traditional business advice, which would be spend 80% and those things that drive revenues and profits today and you're on your core customers and then 20%. Exploring the opposite advice, which is to go all in on everything new, but ignore our core of what we do is sort of very, it goes against all the grains of iconic advice in typical business advice. But, but keep on going.

[B]

Oh no, thank you. Thank you, Scott. And so like, how do you add existing IBM technology? As I cite in the book and I've, I've cited elsewhere, it was hard to get a definition of what Watson exactly was. So you know, back in the day, back in 2015, if you looked at, you know, Google, Microsoft, anybody else that was doing serious machine learning at the time, there was a vocabulary of supervised unsupervised learning within supervised learning, you know, you know, classifiers and regression, et cetera, et cetera. IBM was calling, if you remember this, IBM was calling Watson cognitive computing, which didn't mean anything. I mean it was a marketing phrase. So you had diverging reactions depending if you're an engineer, you didn't take it seriously. If you were like a business person, no offense intended to any business people that are listening. It had wildly inflated expectations of this thing can do anything. So that was tough. And the other thing is, IBM kept trying to tie Watson to only running on IBM cloud to only running on IBM hardware, which made it tough for the acquisitions to do. Even if we wanted to use Watson in our products, it was Tough. And we kept lobbying for this. Look, we're not against it, but just we have to work with our existing products. And what played out years later was showing that we were right because now you have WatsonX in AWS, you have WatsonX in Azure. And so the things that we were pointing out and believe me, they were not well received.

[A]

Well, it was well before people were starting to integrate with any of these AI tools and other types of tools and it was before its time in terms of where it actually got to.

[B]

Yep.

[A]

And so it was. And people couldn't have context to understand it. Now that we're all using ChatGPT and all these other things, we now have a bit of context for where this can go. Even the simple day to day improvements in Google search and other kinds of things where it gives you a quick summary versus having to go through 20 articles. We all have context where this stuff can go today. It's like, it's like selling an app for the iPhone 15 years ago when nobody understood the iPhone platform, you know, and so you couldn't do it. Tell us a second, what are some of the lessons you learned today out of this whole experience? And then tell us the name of the book again. I'm going to ask you that a couple times so everybody hears it. That's the name of the book again.

[B]

So it's the rise and fall of Explorist and IBM Watson Health.

[A]

Thank you. And what are some of the applications today to healthcare tech? What lessons can you learn out of all this stuff that might be applicable today?

[B]

So the. For product development, customers count, customers count and keep iterating. You know, velocity wins and, and we wound up, you know, one of the things I point out in the book was several hitches and dry spells where we're planning and planning and planning and replanning and planning and planning and we're not actually building new products, we're actually talking with our customers. And believe me, if you're not talking with your customers, somebody else will and you will be replaced.

[A]

It's not only just talking to your customers, it's not even that somebody else will be replaced. You end up not building something your customers really want unless you're really, really engaged in talking to them. And I love this concept of keep iterating. We talk about that in startups and different evolutions all the time to iterate till you get a product fit that people actually want and that you guys can produce. So I think that's right on too.

[B]

Yep, it's, it's easy to say, it's sometimes hard to do, but it's, it's, it's advice that never goes out of style. No.

[A]

Thank you very, very much. Thank you. And again, Doug, anything else you learn from this iterate? Velocity, Talk to customers? Any other core messages that people should hear?

[B]

Some people have asked me like, did the failure of Watson Health? It just doesn't make any sense. And so one of the things I try to defend IBM in that I do think it was a good idea. I still applaud IBM for trying, but one of the things that people try to companies that try to take shortcuts through inorganic growth, does that make sense like buying 100%? That is a thing that happens. You know, people forget that YouTube was an acquisition, for example, from, from, you know, from, by Google and you know, Vizio was, was an acquisition from Microsoft. Like acquisitions happen, but the act of integrating those acquisitions into a portfolio, into an integrated portfolio requires organic thinking and organic skills. Does that make sense?

[A]

No, we talk about that all the time. We see companies where organic growth is stalling and organic opportunities are stalling. So they try and buy growth in different ways and it's, it sounds easy, but when the organic engine is going poorly and the core execution is going poorly, sometimes adding on acquisitions doesn't do you any good. If you can't integrate them well anyways, you're not running your business to start with that.

[B]

That's exactly it. And, and that, that was an example of what happened. That was a giant case study there. Yeah, Explorist had organic growth. Vital had organic growth. But unfortunately there weren't enough people that had that kind of A healthcare background and B product development to see that yeah, this is going to require a lot of time and a lot of iteration. And with that I think expectations could have been managed a bit differently. And so again, like I said, I don't fault IBM for trying in healthcare, but yeah, there's definitely a lot wanting in execution.

[A]

No, just fantastic and fascinating. Thank you so much for joining us again, Doug Mel. And the name of the book is the Rise and Fall of Explorers and IBM Watson Health. For those of us that watch this, IBM Watson Health was one of the most promising developments in healthcare. It was gonna be the softball for everything. IBM was coming off the jeopardy thing. IBM was coming off winning the chess match against the great grand masters and they're trying to export into this area and so forth. And it's fascinating story about the rise and fall of a business effort, a business unit that was somewhat before its time. But that's not really why it's failed. It failed for a bunch of other reasons, Doug. Absolutely fascinating. Thank you for joining us today on the Becker Business and the Becker's Healthcare podc. Just a pleasure to visit with you. Thank you very much.

[B]

Thank you, Scott.